Chik's profile - activity

Sorry I am new to ROS. I have an external program that generates an ASCII text file (say 'walk.dat') which consists of two columns of values of linear velocity and angular velocity, say 1000 rows of (a,b). Would like to write a ROS package that can read in the data and execute rostopic pub cmd_vel geometry_msgs/Twist '[a,0,0]' '[0,0,b]' according to the values in the file. I also want to do it by Turtlebot simulator in Gazebo. I really appreciate if someone can teach me step by step from the very beginning (e.g. open a terminal) to the very end (e.g. rosrun something). Thank you very much indeed.

Following the Turtlebot Simulator SLAM tutorial, I copied and pasted the words from the tutorial using Gedit and created the boxes.world, robot.launch, and build_map.launch files. After that, I type roslaunch turtlebot_gazebo_tutorial build_map.launch but the Gazebo window does not show up.

When I type roslaunch turtlebot_gazebo turtlebot_empty_world.launch I can see the virtual turtlebot in the gazebo window, but when I run the build_map.launch the gazebo window does not show up, although from the terminal it seems that the program has been launched successfully. So, what is wrong?

By the way I am using VirtualBox ROS Fuerte by Nootrix. I am not sure if it means a crash of Gazebo in VirtualBox or just I need to source something to bring up the gazebo.

Thank you very much.

Sorry I am new to ROS.

May I ask, is there any navigation stack (for Turtlebot) that allows the robot to reach the goal location (in a known map) and go round some new, small obstacles on-the-way (not shown in the map). Do I need to add some user code in an existing stack?

Thank you very much.

Dear,

I followed the tutorial about building a SLAM map using Turtlebot Simulator. After creating the package, I type:

roslaunch turtlebot_gazebo_tutorial build_map.launch

But it says:

ResourceNotFound: turtlebot_navigation ROS path [0]=/opt/ros/fuerte/share/ros ROS path [1]=/home/viki/sim_workspace/unversioned ROS path [2]=/opt/ros/fuerte/stacks ROS path [3]=/opt/ros/fuerte/share ROS path [4]=/opt/ros/fuerte/share/ros

Would anyone help me? I am using VirtualBox OVA provided by Nootrix. Thank you very much.

David

I used VirtualBox ROS provided by NooTriX. When I launch TurtleBot simulator by typing

roslaunch turtlebot_gazebo turtlebot_empty_world.launch

it keeps showing a warning message

Warning [RaySensor.cc:206] ranges not constructed yet (zero sized)

and as a result, I cannot type anything. Then later, Gazebo closes by itself. Can anyone help me?

Dear,

I just installed VirtualBox Ubuntu 12.04 LTS with ROS Fuerte. After that, I would like to install turtlebot simulator, but found that the tutorial is for electric version. I got stuck at the first step: Configure your Ubuntu repositories to allow "restricted," "universe," and "multiverse" but I cannot see any Software Sources or Software Center. Just wondering if turtlebot can run on Fuerte or not? Or is it the problem of VirtualBox and the LTS version? Thank you very much indeed for your help.

best regards, David

Is there any ROS package that can play songs in iRobot Create of Turtlebot?

Thank you very much.

Just wondering, if there exists any algorithm that builds a map in a topological way (based on landmarks) instead of in a metric way (x,y). Since the gyro/odom has error anyway, it is not necessary to build a metric map. Even humans will not do this. We recognize landmarks to estimate our location, right? Is anyone aware of this kind of research?

I upgraded the Turtlebot laptop to Ubuntu 12.04 and ROS Fuerte. After that, I tried to ssh from my workstation. It gave this error:

@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! @ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY! Someone could be eavesdropping on you right now (man-in-the-middle attack)! It is also possible that a host key has just been changed. The fingerprint for the RSA key sent by the remote host is ca:e6:7b:30:7f:bf:5a:ca:fe:.... Please contact your system administrator. Add correct host key in /home/davidchik/.ssh/known_hosts to get rid of this message. Offending RSA key in /home/davidchik/.ssh/known_hosts:1
remove with: ssh-keygen -f "/home/davidchik/.ssh/known_hosts" -R 172.17... RSA host key for 172.17... has changed and you have requested strict checking. Host key verification failed.

How to reset this key?

Thank you very much.

I was using rviz and roboearth when suddenly the image view went black.

The terminal of openni.launch showed a red sentence "depth image frame id does not match rgb image frame id".

I closed the terminal and opened a new one and then roslaunch openni_launch openni.launch but it said "No devices connected... waiting for devices to be connected".

I reboot the computer a few times, sudo apt-get install ros-electric-openni-kinect, but the problem persists.

Up till this morning the Kinect has been running normally. I really have no idea what happened. Is it due to auto-update? Anyone please help?

Turtlebot laptop: Ubuntu 10.04; ROS Electric

Workstation: Ubuntu 12.04; ROS Fuerte (I have another one in Ubuntu 10.04 and ROS Electric. I tried but the problem remains unsolved.)

The Kinect shows a flashing green light. But when I type lsusb, I saw only one device from Microsoft and one from Chronicity. Is it normal or not?

After roslaunch openni_launch openni.launch, then rosrun dynamic_reconfigure reconfigure_gui, I select /camera/driver, but nothing shows up on the reconfigure window, and the following error is shown on the openni.launch terminal:

[ERROR] [1365395678.038589497]: Client [/reconfigure_gui_13810_1365391644607] wants topic /camera/driver/parameter_descriptions to have datatype/md5sum [dynamic_reconfigure/ConfigDescription/757ce9d44ba8ddd801bb30bc456f946f], but our version has [dynamic_reconfigure/ConfigDescription/5f0f929417f58ac7f30dfdd9e4d40ef2]. Dropping connection.

What is wrong?

Thank you very much for your help.

Have anyone tried using RoboRealm for TurtleBot? From their website it is not clear whether RoboRealm is compatible with ROS or not. Also how does it compare with OpenCV and does it worth buying? Any opinions are welcome.

We are using TurtleBot. There is a laptop directly connected to the TurtleBot. In addition, there is a workstation which connects to the laptop through wireless connection.

According to my understanding, some packages (e.g. teleop) are run from the laptop, while some can be run from the workstation (e.g. rviz). Since we have a powerful workstation, it makes sense to run heavy packages (e.g. opencv, navigation) from the workstation. However, it seems that opencv, amcl, move_base, etc. are now run from the laptop (or am I wrong?)

Hence, my question is, is it possible to run these packages from the workstation? Thank you very much indeed for your help.

The standard TurtleBot mounts the Kinect at <30 cm above the ground. Also there are plates above and below the Kinect. In this case, it is difficult to use the Kinect for object recognition.

Hence I remount it on the top, and also stack up the plates, so the Kinect is now 70 cm above the ground. It seems to have a nice view and can see most objects and people in the office environment. However, it has problems during navigation because of the big blind area in front of its wheel base.

Just wondering, have anyone tried to remount the Kinect at different heights (and angles), and somehow find a most suitable one?

I would like to record my shoes, but they are too big for the current marker template. Can I somehow adjust the parameters and use a bigger marker template to record bigger objects? Thank you very much.

(version 1) Ubuntu 10.04; ROS Electric; rosmake roboearth gives error. This problem cannot be solved even by downgrading the svn.

(version 2) Hence, I upgraded my system to Ubuntu 12.04 LTS; ROS Fuerte; also before rosmake, I install this: sudo apt-get install swi-prolog

Now, the problem is missing knowrob_cad_models but I cannot sudo apt-get install nor rosdep install it.

Today, it is still very challenging to do object recognition in real time while the robot and the object are both moving. Hence, I am thinking of putting some special red/green tags on the target objects or labels. Do you think it will make things better or worse? Also, which ROS package is the most suitable for recognizing the tags? I would like to implement this in TurtleBot with Kinect. Thank you very much for your suggestions.

I would like to use Kinect to recognize 2 objects (book, table), 2 people (adult and child), and 2 postures (stand or sit).

I am confused about OpenCV, OpenNI, PCL, TOD and many things like that. Do I need all of them or just one of them?

Any suggestion for a quick, simple vision package?

In the video "Introducing TurtleBot" in official website, it shows Skeleton Tracking, but what software are they using?

Thank you very much indeed for your help.

(1) When I follow the tutorial to use the 2D Nav Goal in Rviz for TurtleBot, what navigation package is running (e.g. amcl_demo or move_base or what)?

(2) What is the most advanced autonomous navigation package for TurtleBot (e.g. avoid obstacles like PR2)?

Thank you very much.

I am following the tutorial "Visualizing TurtleBot Kinect Data", but I cannot see the data from camera...

roslaunch turtlebot_bringup kinect.launch

it gives:

[ INFO] [1362547950.940390473]: [/openni_camera] Number devices connected: 1
[ INFO] [1362547950.940926034]: [/openni_camera] 1. device on bus 001:10 is a Xbox NUI Camera (2ae) from Microsoft (45e) with serial id 'A00367910023104A'
[ WARN] [1362547950.950968862]: [/openni_camera] device_id is not set! Using first device.
[ INFO] [1362547951.051461572]: [/openni_camera] Opened 'Xbox NUI Camera' on bus 1:10 with serial number 'A00367910023104A'
[ INFO] [1362547951.114948505]: rgb_frame_id = 'camera_rgb_optical_frame' 
[ INFO] [1362547951.127139480]: depth_frame_id = 'camera_depth_optical_frame'

Then on the Workstation, I type:

rosrun rviz rviz

Add camera -> Image Topic -> whatever...

The rviz window shows No Image

And the terminal shows Dropped 100% of messages

Closing rviz, now on the Laptop, I try:

rosrun image_view image_view image:=/camera/rgb/image

It just shows a gray window.

Anyone has any idea what is wrong?

TurtleBot (Kinect, iRobot Create); Ubuntu 10.04; ROS Electric

Please help me!! Thank you very much.

More information: (1) the Kinect shows a flashing green light (not solid light); (2) in rviz, under camera, it says Global Status: Error. Inside it says ... No Transform.

It seems that the visualization tools such as rviz take a lot of computational resources. Actually I just want to watch through the Kinect camera (see what TurtleBot sees) but I don't need other fancy 3d simulation stuff. Is there any simple, high performance package that can do this? Thank you very much.

When I run

rosrun rviz rviz

I received the following

rviz revision number 1.6.7
ogre_tools revision number 1.6.3 
compiled against OGRE version 1.7.3 (Cthugha)
Loading general config from [/home/turtle/.rviz/config]
Loading display config from [/home/turtle/.rviz/display_config]
RTT Preferred Mode is PBuffer.
Texture for pass 0: creating with size 1 x 1 
Segmentation fault

Please help me if anyone knows what is wrong. Thank you very much.

Background: Clearpath TurtleBot with Kinect and iRobot Create; Ubuntu 10.04.